Deep Tech · VC Firm
Lux Capital
Selective quantum exposure via gate-model hardware (Rigetti, 2017 Series A) and the developer-platform layer (Strangeworks, 2018 seed). Not a multi-modality quantum investor; no PsiQuantum, Atom Computing or SandboxAQ holdings.
Investment thesis
Lux Capital is a New York and Menlo Park-based deep-tech VC founded 2000 by Peter Hébert, Josh Wolfe and Robert Paull, with $7B AUM across nine flagship funds. Lux Ventures IX ($1.5B) closed in January 2026, its largest fund. Invests seed to growth across frontier science: AI, bio, defense, space, robotics, materials. Cheques $1-30M.
Quantum exposure: participated in Rigetti Computing's 2017 Series A (a16z-led; gate-model superconducting; public 2022 via SPAC) and Strangeworks' 2018 seed (developer platform). Quantum is one pillar of a broader frontier-computing thesis alongside neuromorphic (Mythic, a DCVC overlap) and photonic AI chips; Lux is not a multi-modality quantum investor and does not hold PsiQuantum, Atom Computing or SandboxAQ.
Value-add: structural dual-use / defense orientation (backed Anduril, Shield AI, Saildrone); strong DOE, DARPA and IC relationships, relevant for quantum founders seeking non-dilutive federal capital; Josh Wolfe's strong narrative platform drives portfolio visibility.
Differentiated as a leading US frontier-computing brand investor with a Rigetti anchor and a proven dual-use defense-tech scaling playbook; quantum remains a concentrated rather than core allocation.
Anti-thesis
The patterns that trigger a fast no with this fund.
- Lux backs weird science with national-security upside: a quantum play with no defense, intelligence or critical-infrastructure angle weakens the thesis
- Its quantum exposure is a small, early gate-model bet (Rigetti, now public) plus a developer platform (Strangeworks); it holds no photonic fault-tolerant (PsiQuantum), neutral-atom (Atom Computing) or PQC (SandboxAQ) names, so the real risk is under-exposure to the modalities now attracting the largest checks
- Lux prizes contrarian, hard-to-replicate physics and IP moats: pure software, middleware or quantum-inspired classical plays rarely fit unless tied to a hardware flywheel
- Josh Wolfe is skeptical of hype cycles and NISQ-era revenue theater: demo-grade fidelity numbers without a credible scaling roadmap to logical qubits will be punished
- Capital intensity is acceptable only with a path to strategic moats (gov contracts, fab partnerships, rare talent); a lean SaaS-style burn profile undercuts the thesis